Inference Model Presentation Using Feature Vectors for User Fit
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems automatically select inference models without considering user-specific use scenes, making it difficult for non-AI familiar users to choose suitable models.
Innovation Solution
An information processing method that identifies and presents suitable inference models to users based on acquired inference target data or keywords, using feature vectors and distance calculations to match models with user needs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automatic model selection is implemented, then model selection speed is improved, but user control over scene-specific suitability is worsened
Solution Approach 1:
The patent introduces an inference target data extraction unit and feature vector generation unit as intermediaries between the automatic model selection system and user needs. These components extract features from user-provided images or keywords, convert them into comparable feature vectors, and use them to automatically identify suitable inference models. This intermediary mechanism enables automatic selection while maintaining scene-specific adaptability by bridging the gap between automated processing and user-specific requirements.
2Loss of time
If automatic model selection is implemented, then selection time is reduced, but ease of use for non-AI users is worsened
Solution Approach 1:
The system enables non-AI users to select suitable inference models through self-service by allowing them to simply input inference target images or keywords. The system automatically extracts features from these inputs, generates feature vectors, compares them with pre-stored model feature vectors, and identifies suitable models without requiring users to have AI expertise. This self-service approach significantly reduces selection time while maintaining ease of use for non-AI users.
3Adaptability or versatility
If manual model selection is implemented, then user control is improved, but selection cost and time are worsened
Solution Approach 1:
The patent implements preliminary action by pre-storing feature vectors for multiple inference models before the actual selection process. When a user needs to select a model, the system has already prepared the reference feature vectors, so it only needs to extract features from the user's inference target data and perform comparison. This preliminary preparation significantly reduces the time and computational cost required during the actual selection process while maintaining user control over the selection outcome.
Data Source
AI summary
A model presentation device includes a keyword acquisition part that acquires at least one piece of inference target data, an identification part that identifies at least one inference model according to the at least one piece of inference target data from among a plurality of inference models that output an inference result using the inference target data as an input, a presentation screen creation part that creates a presentation screen for presenting the identified at least one inference model to a user, and a display part that outputs the created presentation screen.